Content Fans

Category

AI News & Trends

967 articles • Page 12 of 65

AI-generated code triples remediation time, costs $4.88M per breach

AI-generated code triples remediation time, costs $4.88M per breach

AI-generated code may create security gaps that are hard to spot and fix. When problems happen, fixing the damage reportedly takes about three times longer than with human-written code and costs around $4.88 million per breach on average. Studies suggest that almost half of AI-generated code samples may have security flaws, and big companies have seen a big rise in security issues after using AI tools. Experts recommend strict access controls and regular checks to lower risks, but teams often spend a lot of time just figuring out what the AI was supposed to do before fixing any problems.

AI Coding Agents Shift Engineers to Oversight in 2026

AI Coding Agents Shift Engineers to Oversight in 2026

By mid-2026, coding agents may be handling most of the routine coding work, such as writing code, testing, and making pull requests, while human engineers focus on supervision and review. Reports suggest agents like OpenAI Codex, Claude Code, Cursor Composer, and GitHub Copilot Agent Mode now manage full workflows, including debugging and documentation. Teams choose between using one agent for simple tasks or several agents in parallel for bigger, modular tasks, depending on the job and oversight needs. No single tool appears to lead in all benchmarks, and the best option seems to depend on the kind of task. Forecasts suggest engineers might spend more time on oversight and coordination, though some experts warn that progress may be slower than hoped.

AI Agents Reshape SaaS, Threaten Seat Licenses by 2026

AI Agents Reshape SaaS, Threaten Seat Licenses by 2026

AI agents may change how SaaS companies price and deliver their products by 2026. Reports suggest that traditional seat licenses could become less important as AI agents get better at doing work for many people, not just guiding them. Experts believe vendors are slowly adding AI features to test what works, which might help them avoid big risks. Some studies predict many enterprise apps will have AI agents by 2026, and companies focused only on selling many seats may struggle. Buyers seem to value practical AI features like automation and good data handling, so vendors who offer these may adapt better to changes.

Anthropic updates Claude Opus 4.8: Faster, cheaper, and safer AI

Anthropic updates Claude Opus 4.8: Faster, cheaper, and safer AI

Anthropic has released Claude Opus 4.8, which it claims may be faster, cheaper, and safer for tasks like coding, finance, and knowledge work. Early feedback and benchmarks suggest it performs well in some areas, such as bug fixing and long-context reasoning, but might not be as strong at command-line automation compared to rivals. The company reports lower costs, quicker responses, and fewer cases of risky behavior than previous models. Some early users have seen better results in tasks like legal review and report drafting. However, results appear mixed depending on the task, and industry voices suggest it may be a strong choice for certain uses but not all.

Israel Warns Iran Uses AI to Polish Disinformation, Recruitment

Israel Warns Iran Uses AI to Polish Disinformation, Recruitment

Israel's National Cyber Directorate warns that Iranian hackers may be using AI to make disinformation and recruitment messages more convincing and harder to detect. These AI-powered efforts appear to include mass text messages and fake social media profiles, sometimes using very natural-sounding Hebrew, which suggests generative AI is involved. Officials say these campaigns often try to cause panic or recruit Israelis to share information. Experts believe that AI allows these groups to quickly change their messages for different audiences, making operations cheaper and faster. Israel is responding by advising the public to verify alerts and is running drills to prepare for both technical and psychological attacks.

MiniMax M3 Unveils 1M-Token Context for Enterprise AI Agents

MiniMax M3 Unveils 1M-Token Context for Enterprise AI Agents

MiniMax M3 is an open AI model that may help enterprises handle large and complex tasks, like coding and video analysis, by supporting up to 1 million tokens of context. Benchmarks suggest it performs well on long and multi-step tasks, possibly matching top closed models on some measures. Companies appear to be choosing open models like M3 for lower costs and more control, especially where data privacy or custom tuning is needed. However, some uncertainty remains, as closed models may still lead in special features and reliability. The market now seems to focus on which models can handle long contexts, different data types, and complex agent work.

Google Deepens Workspace AI Integration, Raising Privacy Concerns

Google Deepens Workspace AI Integration, Raising Privacy Concerns

Google is making its Gemini AI assistant a bigger part of Workspace apps like email and documents, which may raise new privacy and security concerns. Experts warn that if users have broad access, the AI might reveal sensitive information more easily, and AI-generated summaries could create records that are hard to delete. Most companies still use AI in limited ways and focus on security reviews before using it more widely. There may be skill gaps, as many workers know about AI tools but might not have formal training to use or check them. Early reports suggest that careful settings and strong permissions could help avoid privacy problems as AI gets used more often at work.

Google's AI Mode Exceeds 1 Billion Users, Reshaping Search

Google's AI Mode Exceeds 1 Billion Users, Reshaping Search

Google's new AI Mode in Search may be changing how people use the internet, with over 1 billion monthly users since its launch. Instead of just showing links, AI Mode gives chat-like answers and suggests follow-up questions, though users can still use the regular web tab. Reports suggest that most users now get their answers without clicking on outside links, which may be causing less traffic to other websites. There is some concern that this could hurt publishers and change how sites try to appear in search results. While competitors are trying similar things, none seem to have as many users as Google's AI Mode.

OpenAI Launches DeployCo: $4 Billion Unit Embeds AI Engineers for Enterprises

OpenAI Launches DeployCo: $4 Billion Unit Embeds AI Engineers for Enterprises

OpenAI has announced DeployCo, a new company with over $4 billion in funding, to help businesses use AI by embedding its engineers directly into client companies. These engineers, called forward-deployed engineers (FDEs), work closely inside businesses to set up and support AI systems, which may make it harder for clients to switch to other vendors later. Reports suggest this setup helps companies solve problems faster and may lead to higher customer retention, but large-scale results are not yet published. DeployCo will start with about 150 FDEs from OpenAI's acquisition of Tomoro, and investors may gain special data rights as part of these partnerships. The shift to these deep deployment partnerships may indicate that businesses now value real results from AI over just test scores.

Palo Alto Networks tests Claude Mythos for vulnerability detection, finds 26 CVEs

Palo Alto Networks tests Claude Mythos for vulnerability detection, finds 26 CVEs

Palo Alto Networks tested the Claude Mythos AI tool for finding software vulnerabilities and found 26 possible issues, though none appear to have been exploited. The results suggest that while AI can quickly surface meaningful security problems, each discovery still needs human review and might add extra costs. Real-world buyers may face high expenses once free usage credits run out. There may also be risks if the tools are not properly managed, like leaking sensitive code. Organizations might benefit from these tools if they have large, complex systems, but smaller teams may find traditional methods work well enough at a lower cost.

Google: AI shrinks cyberattack cost, boosts volume in 2026

Google: AI shrinks cyberattack cost, boosts volume in 2026

Google's report suggests AI is making cyberattacks cheaper and more frequent. Attackers may now use AI to quickly find security flaws, create convincing phishing emails, and adjust malware to avoid detection. Early cases show AI can help hackers work much faster than before, though defenders are starting to use AI tools too. Experts warn that organizations who delay updating their defenses might struggle against these fast, automated attacks. The report says the best defense may be combining AI detection with human oversight and strong security practices.

NVIDIA Unveils RTX Spark AI PCs With 1 Petaflop Power

NVIDIA Unveils RTX Spark AI PCs With 1 Petaflop Power

NVIDIA has introduced a new type of AI PC called RTX Spark, which may reach up to 1 petaflop of computing power in devices the size of a typical gaming laptop. These systems combine a Grace CPU with a Blackwell GPU and can support up to 128 GB of unified memory, which might allow users to run much larger AI models on their own device. RTX Spark comes with special security features like session isolation and strict permission controls through the NVIDIA OpenShell software. Release dates are still vague, but NVIDIA says these PCs will be available in the fall, with details about pricing and full performance still to come.

Open-weight models integrate agentic features, challenge proprietary AI in 2026

Open-weight models integrate agentic features, challenge proprietary AI in 2026

Open-weight AI models are adding more advanced features like long-context processing, tool use, and handling different data types, and they may be catching up to proprietary AI models. Early results, such as those from MiniMax M3, suggest that open models are becoming better at real tasks, especially coding. Many companies seem to prefer open models because they can control, customize, and manage costs more easily. However, most companies are still only partly using agentic AI in their daily work, with strong rules and monitoring needed before bigger rollouts. Experts suggest that the main competition now may be about control, reliability, and cost, not just accuracy.

Hershey invests $250M to accelerate supply chain automation by 2026

Hershey invests $250M to accelerate supply chain automation by 2026

Hershey plans to invest $250 million by 2026 to make its supply chain more automated and digitally connected. The company expects this may save about $300 million a year, with a large part coming from better supply chain productivity. Hershey's approach suggests other companies should first standardize processes, then connect real-time data, and only later expand automation. The roadmap points to using key technology systems and careful pilot projects. Some surveys warn that digital projects might fail if workers lack the right skills, so training and upskilling appear important for success.